Characterization of Cyanobacterial Hydrocarbon Composition and Distribution of Biosynthetic Pathways R. Cameron Coates 1 , Sheila Podell 1 , Anton Korobeynikov 3,4 , Alla Lapidus 3,5 , Pavel Pevzner 3,6 , David H. Sherman 7 , Eric E. Allen 1 , Lena Gerwick 1 , William H. Gerwick 1,2 * 1 Center for Marine Biotechnology and Biomedicine, Scripps Institution of Oceanography, University of California San Diego, La Jolla, California, United States of America, 2 Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California San Diego, La Jolla, California, United States of America, 3 Algorithmic Biology Laboratory, St. Petersburg Academic University, Russian Academy of Sciences, St. Petersburg, Russia, 4 Department of Mathematics and Mechanics, St. Petersburg State University, St. Petersburg, Russia, 5 Theodosius Dobzhansky Center for Genome Bionformatics, St. Petersburg State University, St. Petersburg, Russia, 6 Department of Computer Science and Engineering, University of California San Diego, La Jolla, California, United States of America, 7 Life Sciences Institute and Department of Medical Chemistry, University of Michigan, Ann Arbor, Michigan, United States of America Abstract Cyanobacteria possess the unique capacity to naturally produce hydrocarbons from fatty acids. Hydrocarbon compositions of thirty-two strains of cyanobacteria were characterized to reveal novel structural features and insights into hydrocarbon biosynthesis in cyanobacteria. This investigation revealed new double bond (2- and 3-heptadecene) and methyl group positions (3-, 4- and 5-methylheptadecane) for a variety of strains. Additionally, results from this study and literature reports indicate that hydrocarbon production is a universal phenomenon in cyanobacteria. All cyanobacteria possess the capacity to produce hydrocarbons from fatty acids yet not all accomplish this through the same metabolic pathway. One pathway comprises a two-step conversion of fatty acids first to fatty aldehydes and then alkanes that involves a fatty acyl ACP reductase (FAAR) and aldehyde deformylating oxygenase (ADO). The second involves a polyketide synthase (PKS) pathway that first elongates the acyl chain followed by decarboxylation to produce a terminal alkene (olefin synthase, OLS). Sixty-one strains possessing the FAAR/ADO pathway and twelve strains possessing the OLS pathway were newly identified through bioinformatic analyses. Strains possessing the OLS pathway formed a cohesive phylogenetic clade with the exception of three Moorea strains and Leptolyngbya sp. PCC 6406 which may have acquired the OLS pathway via horizontal gene transfer. Hydrocarbon pathways were identified in one-hundred-forty-two strains of cyanobacteria over a broad phylogenetic range and there were no instances where both the FAAR/ADO and the OLS pathways were found together in the same genome, suggesting an unknown selective pressure maintains one or the other pathway, but not both. Citation: Coates RC, Podell S, Korobeynikov A, Lapidus A, Pevzner P, et al. (2014) Characterization of Cyanobacterial Hydrocarbon Composition and Distribution of Biosynthetic Pathways. PLoS ONE 9(1): e85140. doi:10.1371/journal.pone.0085140 Editor: Bing Xu, Brandeis University, United States of America Received July 12, 2013; Accepted November 22, 2013; Published January 27, 2014 Copyright: ß 2014 Coates et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Funding for this work came from CEC 500-10-039, NIH CA108874, NIH GM067550, NSF MCB-1149552 (EEA). DHS acknowledges support from the International Cooperative Biodiversity Group TW007404. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected]Introduction Cyanobacteria are a diverse group of photosynthetic bacteria that have evolved a remarkable array of adaptive traits including oxygenic photosynthesis, N 2 fixation, a wide morphological diversity, extensive secondary metabolite biosynthetic capacity, and a range of symbiotic relationships with other organisms. Cyanobacteria are estimated to contribute 30% of Earth’s annual oxygen production and play a major role in biogeochemical cycles [1]. One trait less well characterized and potentially of great societal importance is their universal ability to produce long chain hydrocarbons. First recognition of this latter trait resulted from investigations in the 1960’s [2–4] and was of importance in the context of identifying the origin of hydrocarbons found in sedimentary and oil deposits. In recent years, there has been a growing recognition of the negative environmental impacts of continued fossil fuel use, as well as an ever increasing worldwide energy demand, and these facts have combined to increase interest in developing sustainable biofuels such as hydrocarbons from cyanobacteria [5,6]. Many sources of renewable energy can be envisioned to help meet society’s demand for electrical power, however, there remains an acute need for low cost liquid fuels, and particularly gasoline, diesel and jet fuel [6]. Diesel and jet fuel quality are measured using cetane values whereas octane ratings are the principal measure of gasoline quality. Quality of the fuel is thus largely determined by the structure of the hydrocarbons in the fuel, and the type of fuel being considered. Long, straight-chain saturated hydrocarbons exhibit higher cetane ratings whereas short highly branched molecules (i.e. isooctane) exhibit superior octane ratings [7,8]. It is important to consider these structural characteristics when assessing the applicability of various biofuel molecules to different engine applications. In this context, the commonly observed hydrocar- bons of cyanobacteria, including heptadecane (cetane rating: 105) and methylheptadecane (cetane rating: 66), are promising candidates for diesel fuel applications (normal cetane rating: 40– 55) [5,7,8]. PLOS ONE | www.plosone.org 1 January 2014 | Volume 9 | Issue 1 | e85140
12
Embed
Characterization of Cyanobacterial Hydrocarbon Composition ......Hydrocarbons produced via fatty acid pathways have also been described in various eukaryotic organisms including the
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Characterization of Cyanobacterial HydrocarbonComposition and Distribution of Biosynthetic PathwaysR. Cameron Coates1, Sheila Podell1, Anton Korobeynikov3,4, Alla Lapidus3,5, Pavel Pevzner3,6,
David H. Sherman7, Eric E. Allen1, Lena Gerwick1, William H. Gerwick1,2*
1 Center for Marine Biotechnology and Biomedicine, Scripps Institution of Oceanography, University of California San Diego, La Jolla, California, United States of America,
2 Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California San Diego, La Jolla, California, United States of America, 3 Algorithmic Biology
Laboratory, St. Petersburg Academic University, Russian Academy of Sciences, St. Petersburg, Russia, 4 Department of Mathematics and Mechanics, St. Petersburg State
University, St. Petersburg, Russia, 5 Theodosius Dobzhansky Center for Genome Bionformatics, St. Petersburg State University, St. Petersburg, Russia, 6 Department of
Computer Science and Engineering, University of California San Diego, La Jolla, California, United States of America, 7 Life Sciences Institute and Department of Medical
Chemistry, University of Michigan, Ann Arbor, Michigan, United States of America
Abstract
Cyanobacteria possess the unique capacity to naturally produce hydrocarbons from fatty acids. Hydrocarbon compositionsof thirty-two strains of cyanobacteria were characterized to reveal novel structural features and insights into hydrocarbonbiosynthesis in cyanobacteria. This investigation revealed new double bond (2- and 3-heptadecene) and methyl grouppositions (3-, 4- and 5-methylheptadecane) for a variety of strains. Additionally, results from this study and literature reportsindicate that hydrocarbon production is a universal phenomenon in cyanobacteria. All cyanobacteria possess the capacityto produce hydrocarbons from fatty acids yet not all accomplish this through the same metabolic pathway. One pathwaycomprises a two-step conversion of fatty acids first to fatty aldehydes and then alkanes that involves a fatty acyl ACPreductase (FAAR) and aldehyde deformylating oxygenase (ADO). The second involves a polyketide synthase (PKS) pathwaythat first elongates the acyl chain followed by decarboxylation to produce a terminal alkene (olefin synthase, OLS). Sixty-onestrains possessing the FAAR/ADO pathway and twelve strains possessing the OLS pathway were newly identified throughbioinformatic analyses. Strains possessing the OLS pathway formed a cohesive phylogenetic clade with the exception ofthree Moorea strains and Leptolyngbya sp. PCC 6406 which may have acquired the OLS pathway via horizontal gene transfer.Hydrocarbon pathways were identified in one-hundred-forty-two strains of cyanobacteria over a broad phylogenetic rangeand there were no instances where both the FAAR/ADO and the OLS pathways were found together in the same genome,suggesting an unknown selective pressure maintains one or the other pathway, but not both.
Citation: Coates RC, Podell S, Korobeynikov A, Lapidus A, Pevzner P, et al. (2014) Characterization of Cyanobacterial Hydrocarbon Composition and Distributionof Biosynthetic Pathways. PLoS ONE 9(1): e85140. doi:10.1371/journal.pone.0085140
Editor: Bing Xu, Brandeis University, United States of America
Received July 12, 2013; Accepted November 22, 2013; Published January 27, 2014
Copyright: � 2014 Coates et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Funding for this work came from CEC 500-10-039, NIH CA108874, NIH GM067550, NSF MCB-1149552 (EEA). DHS acknowledges support from theInternational Cooperative Biodiversity Group TW007404. The funders had no role in study design, data collection and analysis, decision to publish, or preparationof the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
8, Okeania sp. PAC-18-Feb-10-1.1 were isolated from field
collections and are maintained among the Gerwick Laboratory
Figure 1. Hydrocarbon biosynthetic pathways in cyanobacteria. A) The Fatty Acyl-ACP Reductase (FAAR)/Aldehyde DeformylatingOxygenase (ADO) involves first a reduction of a fatty acyl substrate to a fatty aldehyde followed by an oxidative conversion to an alkane with therelease of formate [17]. The OLS (olefin producing) pathway involves a polyketide synthase that first elongates a fatty acyl-CoA by two carbons frommalonyl-CoA via ketosynthase (KS) and acyl transferase (AT) domains followed by reduction to the b-hydroxyacid by a ketoreductase (KR). The STactivates the b-hydroxy group via sulfonation, and then the thioesterase (TE) acts on this substrate to catalyze decarboxylation and loss of sulfate toform the terminal alkene.doi:10.1371/journal.pone.0085140.g001
Cyanobacterial Hydrocarbons
PLOS ONE | www.plosone.org 3 January 2014 | Volume 9 | Issue 1 | e85140
Cyanobacterial Hydrocarbons
PLOS ONE | www.plosone.org 4 January 2014 | Volume 9 | Issue 1 | e85140
culture collection as described in [24,34,35]. Planktothrix agardhii
NIVA-CYA 168 was provided by Rainer Kurmayer at the
Austrian Academy of Sciences Institute for Limnology. Haplosiphon
welwitchii IC-52-3 and Westiella intricata HT-29-1 were provided by
Thomas Hemscheidt of University of Hawaii via Melinda Micallef
and Michelle Moffitt. The remaining strains were purchased from
the Pasteur Culture Collection of Cyanobacteria (PCC) or the
American Type Culture Collection (ATCC) as identified in the
text.
Culture ConditionsCyanobacterial strains were grown in 2.8 L Fernbach flasks
under 16:8 day: night light cycle between 20–60 mE/m2/sec at a
constant temperature at 20uC, 25uC, or 28uC in BG-11, SWBG-
11, or ASNIII. Cultures were shaken continuously (80 rpm) or
grown statically between 14 and 35 days depending upon strain
growth rates. Cultures (1L) were harvested via centrifugation at
4000 rpm for 15 min in 500 mL conical containers and combined
(using 0.5 M ammonium formate to remove salts for marine
strains and deionized water for freshwater strains) into 50 mL
Falcon tubes and subsequently centrifuged again to yield a packed
pellet that was frozen and dried for extraction and analysis. For
filamentous strains that were not amenable to centrifugation,
filaments were removed from the media using forceps (rinsed using
0.5 M ammonium formate to remove salts for marine strains), and
frozen for drying, extraction and analysis. All biomass was
lyophilized for at least 24 h.
Extraction and Structural AnalysisDried biomass was ground using a mortar and pestle and
weighed. Biomass was extracted using 5 mL of 100% hexanes or
2:1 dichloromethane:methanol (DCM:MeOH) followed by 20 sec
of sonication. The extract was filtered using Whatman GF/F and
the residual biomass re-extracted two additional times using the
same method followed by a 10 mL wash with hexanes or 2:1
DCM:MeOH. Octadecane (Fluka-74691) was added to each
extract after initial solvent addition as an internal standard for
quantitation. Octadecane was added at approximately 0.1% of the
initial dry biomass. DCM:MeOH (2:1) crude extracts were
fractionated using a normal phase 500 mg Bonna-Agela Cleanert
silica SPE column with collection of the first fraction which eluted
with 100% hexanes. Extracts were dried under N2 gas. A
comparison between hexane and DCM:MeOH extraction meth-
ods for three strains (Anabaena (Nostoc) PCC 7120, S. elongatus PCC
7942, Synechococcus sp. PCC7002) found that yields and composition
of hydrocarbons were not significantly different (data not shown).
To characterize potential fatty acid substrates for hydrocarbon
biosynthesis, a fatty acid analysis was performed on crude extracts
(2:1 DCM:MeOH) of Anabaena (Nostoc) sp. PCC7120, M. producens
3L, and Synechococcus sp. PCC 7002. Fatty acid methyl esters
(FAMEs) were produced by transesterification by adding 3 mL of
4% H2SO4 (in MeOH) to at least 0.1 mg of a crude extract.
Samples were then stirred and incubated at 110uC for 1 h. Four
mL of H2O and 3 mL of hexanes were then added to the sample.
After vortexing for 30 sec and centrifugation at 2500 rpm for
3 min the hexanes layer (top) was removed and dried in a pre-
weighed vial for GC-MS analysis. FAME preparation was
followed by dimethyl disulfide (DMDS) derivitization for deter-
mination of double bond positions [36].
Each extract was resuspended to a concentration of 100 mg/mL
in hexanes and 1 mL was analyzed by gas chromatography mass
spectrometry (GC-MS) using a Thermo Trace GC-DSQ instru-
ment equipped with an Agilent DB5-ms column (30 m, ID: 0.25,
Film: 0.25 mm). Helium (constant flow 1 mL/min) was used as the
carrier gas. The inlet temperature was 240uC and the following
temperature program was applied: 40uC for 1 min with an
increase of 4.5uC/min to 250uC for 10 min. Data were acquired
and processed with the Thermo Xcaliber software. Hydrocarbons
were determined using a combination of mass fragmentation
patterns, retention time and comparison to authentic standards
when available [heptadecane (Fluka-51578), 1-heptadecene (TCI-
S0347), 7-methylheptadecane (kindly provided by Dieter Enders
and Wolfgang Bettray, RWTH, Aachen University)], or published
mass spectra and the NIST mass spectral library for Xcaliber
(2005) when not available. Double bond positions were confirmed
for all alkenes and unsaturated fatty acids using the DMDS
method [36].
GC-MS detector response factors for heptadecane, 1-heptade-
cene, and 7-methylheptadecane were determined in comparison
with the octadecane standard by creation of standard curves.
Standard curves for hydrocarbons with authentic standards were
verified using the low mass common to all hydrocarbons analyzed
via GC-MS (57 m/z). Hydrocarbon concentration was calculated
using hydrocarbon peak area compared to the internal standard
(octadecane) peak area using 57 m/z. Percent dry weight was
calculated as an average of three biological replicates. Statistical
analysis of the hydrocarbon yields between the OLS and FAAR/
ADO pathway were completed using a Mann-Whitney-Wilcox
Test.
Genome Sequencing and Bioinformatic AnalysisThe sequenced genomes of M. bouillonii PNG5-198, M. producens
JHB, and cf. Phormidium sp. ISB 3/Nov/94-8 were generated at
either the Genomic Center at The Scripps Research Institute or at
the University of Michigan using Illumina technology [37].
Genomes were corrected with Quake and assembled using
SPAdes 2.4 [38,39]. Contigs were binned by GC content to
remove non-cyanobacterial DNA sequences.
Identification of cyanobacterial hydrocarbon pathways was
accomplished using blastn searches against newly sequenced
genomes and blastp searches using representative genes from
each pathway (FAAR/ADO and OLS) against publicly available
cyanobacterial genomes (131 total) from GenBank [40] and the
Joint Genome Institute (JGI) Integrated Microbial Genomes
(IMG) database (Version 4, [41]. Draft genome sequences for
M. bouillonii PNG5-198, M. producens JHB, cf. Phormidium sp. ISB 3/
Nov/94-8, P. agardii NIVA-CYA 126/8 (provided by Rainer
Kurmayer, bio project accession number: PRJNA163669), H.
Figure 2. Cyanobacterial 16S rRNA phylogeny and hydrocarbon pathway distribution. The 16S rRNA phylogeny of publicly availablegenome sequenced cyanobacteria (128) and additional strains investigated in this study (14) including G. violaceus PCC 7421 as the outgroup. Bluestrain names indicate strains possessing the FAAR/ADO pathway and red indicates those with the OLS pathway. Purple strain names indicate a strainthat does not have a genome sequence and therefore pathway presence cannot yet be verified. Cyanobacterial subdivisions are labeled usingcolored branches following the key in the upper left: (1) Subdivision 1. Uniceullular (Formerly Chroococcales), Subdivision II. Baeocystous (FormerlyPleurocapsales), Subdivision III. Filamentous (Formerly Oscillatoriales), Subdivision IV. Heterocystous (Formerly Nostocales), Subdivision V. Ramified orTrue Branching (Formerly Stigonematales). The clade indicated as ‘‘A’’ represents the main clade of cyanobacteria with the OLS pathway while clade‘‘B’’ indicates the Moorea strains and clade ‘‘C’’ indicates the clade containing Leptolyngbya sp. PCC 6404. Baysian posterior probabilities are displayedat nodes (o = posterior support,0.5).doi:10.1371/journal.pone.0085140.g002
Cyanobacterial Hydrocarbons
PLOS ONE | www.plosone.org 5 January 2014 | Volume 9 | Issue 1 | e85140
welwitschii IC-52-3 and W. intricata HT-29-1 (provided by Melinda
Micallef and Michelle Moffitt), and Leptolyngbya sp. BL0902
(provided by Arnaud Taton, Susan Golden and James Golden)
were also used. The sulfotransferase domain in the OLS pathway
identified in Synechococus sp. PCC 7002 was used for this search
because this is a distinctive enzymatic step in terminal alkene
biosynthesis [22]. The FAAR and ADO enzymes from S. elongatus
PCC 7942 were used as representative sequences in the search for
FAAR/ADO pathways. Searches for FAAR/ADO pathway
enzymes were also performed using hidden Markov model protein
family patterns TIGR04058 (aldehyde-forming long-chain fatty
acyl-ACP reductase) and TIGR04059 (long-chain fatty aldehyde
decarbonylase) from the TIGRFAMs database [42].
Phylogenetic AnalysisTo establish and compare the phylogenetic distribution of
hydrocarbon biosynthetic pathways in cyanobacteria, phylogenies
were constructed using the 16S rRNA gene sequences from all
cyanobacterial strains for which genome sequence data is
publically available (128) as well as all additional cyanobacteria
investigated in this study (16). Due to incomplete or partial
sequences, cf. Phormidium sp. ISB 3/Nov/94-8, W. intricata HT-29-
1, and Synechococcus sp. CB0205 were omitted from these analysis.
Pleurocapsa sp. PCC 7320, Cyanobacterium sp. JSC-1 as well as
Synechococcus sp. CB0101 were not included in the phylogenetic
analyses because 16S rRNA sequences are not currently available
for these strains. Alignments of the 16S rRNA sequences were
completed using MAFFT and trees were generated using both
PCC 6406 is from a distinct evolutionary lineage from other
OLS-containing cyanobacteria yet appears to possess this same
metabolic pathway.
Pathway EvolutionAn alignment of DNA sequences for all known OLS pathways
was constructed with annotations for open reading frames (Figure
S4). This alignment also includes the CurM domain of the curacin
A biosynthetic pathway, and as expected, shows that CurM does
not contain the fatty acyl ACP ligase (FAAL) domain or the first
acyl carrier protein (ACP) found in the OLS pathway (Figure S4).
An alignment of the amino acid sequences of all 17 of the known
OLS pathways and CurM shows that all of the OLS pathways
contain the same domains and domain architecture (Figure S5). A
phylogenetic tree of all OLS pathways was created to investigate
their evolutionary relationships (Figure S6). This phylogenetic tree
shows a similar topology with that produced from the correspond-
ing 16S rRNA sequences (Figure 2) with the exception of
Leptolyngbya sp. PCC 6406. In this latter case, the OLS sequence
clades with those from the Moorea strains and CurM. A
phylogenetic tree was generated for all available ADO genes as
well; however, evolutionary relationships and topological compar-
isons to the 16S phylogenetic tree were prohibited due to poor
bootstrap support (data not shown). Significantly, none of the
genome sequenced cyanobacteria appear to contain both the
FAAR/ADO pathway as well as the OLS pathway. Additionally,
no pseudogenes with homology to either of the genes in the
FAAR/ADO pathway were detected in OLS-containing cyano-
bacteria, and similarly, no pseudogenes were detected for the OLS
pathway in strains containing the FAAR/ADO pathway.
Discussion
A striking general conclusion that emerges from review of
previous studies along with the results of this study is that
hydrocarbon production is a universal phenomenon among
cyanobacteria (Figure 2–4 and Table S1). However, hydrocarbon
production in cyanobacteria appears to be derived from at least
two very different pathways. Moreover, the specific structural
features of the hydrocarbons reflect which pathway is present;
saturated alkanes are found in strains with the FAAR/ADO
Figure 3. Quantitative yields of hydrocarbons as percent dry weight of biomass from 20 cyanobacteria displayed by phylogeneticrelationship and pathway distribution. Specific hydrocarbons for each strain are color coded and stacked to depict the overall quantitative yield.Standard error bars are given for each hydrocarbon and each strain. Quantitative hydrocarbon yields ranged from 0.024%60.01% in Cyanothece sp.PCC 7425 to 0.262%60.01% in Pleurocapsa sp. PCC 7516. Blue strain names indicate strains possessing the FAAR/ADO pathway and red indicatesthose with the OLS pathway. Purple strain names indicate a strain that does not have a genome sequence and therefore the pathway type isunknown. To the left of the figure, a 16S rRNA phylogenetic tree (Maximum Likelihood, see Figure S1 for complete tree) is presented for all 20displayed strains. Branch tips are aligned to corresponding strain names except for Westiella intricata HT-29-1 for which no 16S rRNA sequence isavailable. Vertical connection lengths were modified to accommodate the location of W. intricata HT-29-1 in the table.doi:10.1371/journal.pone.0085140.g003
Cyanobacterial Hydrocarbons
PLOS ONE | www.plosone.org 7 January 2014 | Volume 9 | Issue 1 | e85140
pathway and terminal alkene containing-hydrocarbons are found
only in OLS containing strains. As reported previously, heptade-
cane is the most commonly observed hydrocarbon in cyanobac-
teria followed by heptadecene, pentadecane and 7-methylhepta-
decane (Table S1) [2,5,44]. This observation is consistent with
octadecanoic acid (FAAR/ADO) or hexadecanoic acid (OLS)
precursor fatty acids, and these along with a variety of unsaturated
derivatives are the most common fatty acids in cyanobacteria [45].
Hydrocarbon CompositionMost of the strains investigated in this study had never been
previously characterized for their hydrocarbon composition (22 of
32). This investigation found a wide variation in the content of
hydrocarbons between these cyanobacterial strains with a range of
only 0.024%60.01% in Cyanothece sp. PCC 7425 to
0.262%60.01% in Pleurocapsa sp. PCC 7516 (Figure 3). These
results expand the previously reported range of natural cyano-
bacterial hydrocarbon yields that was between 0.025–0.12% dry
weight [2,4]. Strains possessing the OLS pathway appear to have
significantly higher hydrocarbon yields (0.173%60.032) than
strains with the FAAR/ADO pathway (0.070%60.008) (p-
value = 0.0002).
Branched alkanes have been used as a biomarker for
cyanobacteria, and consistent with this, they appear to be widely
distributed across cyanobacterial phylogeny (Figure 2, Table S1)
[2,44]. Branched alkanes have been observed mostly in filamen-
tous cyanobacteria but there are a few reports of branched
hydrocarbons in unicellular strains (Anacystis nidulans, Anacystis
cyanea, and Chrocococcus turgidus) [2,30,33]. Expanding this distribu-
tion, we show for the first time that the unicellular cyanobacterium
G. violaceus PCC 7421 also produces 7-methylheptadecane. This
study also expanded the known structural diversity of cyanobac-
terial hydrocarbons to include additional double bond positions (2-
and 3-heptadecenes) and methyl group positions (3-, 4- and 5-
methylheptadecanes). Overall, branched alkanes were observed in
7 of the 32 strains examined in this investigation. Branched alkane
biosynthesis was previously investigated in cyanobacteria and the
pendant methyl group found to be derived from S-adenosylme-
thionine (SAM) through a methyltransferase reaction [46,47].
However, the methyltransferase involved in this pathway has yet to
be identified and characterized. Han et al. [2] and Fehler & Light
[46] used radiolabeled substrates to verify that vaccenic acid (11-
octadecenoic acid) is the likely precursor to 7- or 8-methylhepta-
decane in Nostoc muscorum. We also observed 11-octadecenoic acid
in our fatty acid analysis of Anabaena (Nostoc) sp. PCC 7120, thus
confirming the possibility that this fatty acid is the precursor to
these branched hydrocarbons in this strain (Figure 5).
All three Moorea strains as well as D. incrassata PCC 7326
produced the unique alkenes 2- and 3-heptadecene. Two possible
pathways may be responsible for the production of these observed
alkenes. First, a desaturase could act upon a fatty acid substrate to
produce double bonds at the v-2 or v-3 position. The unsaturated
Figure 4. Hydrocarbon composition expressed as a percentage of total hydrocarbons for 32 strains of cyanobacteria displayed byphylogenetic relationship and pathway distribution. Percentages are displayed as mean percentage between three replicates except for thoseindicated with an asterisk for strains that were characterized using a single sample. Blue strain names indicate strains possessing the FAAR/ADOpathway and red indicates those with the OLS pathway. Purple strain names indicate a strain that does not have a genome sequence and thereforethe pathway type is unknown. To the left of the figure, a 16S rRNA phylogenetic tree (Maximum Likelihood, see Figure S2 for complete tree) ispresented for all 32 displayed strains. Branch tips are aligned to the corresponding strain names. Branches corresponding to W. intricata HT-29-1, cf.Phormidium sp. ISB 3/Nov/94-8 and Pleurocapsa sp. PCC 7320 are not shown because 16s rRNA sequences are not available for these strains. Verticalconnection lengths were modified to accommodate the location of these three strains in the table.doi:10.1371/journal.pone.0085140.g004
Cyanobacterial Hydrocarbons
PLOS ONE | www.plosone.org 8 January 2014 | Volume 9 | Issue 1 | e85140
fatty acid could then undergo elongation and decarboxylation via
the OLS pathway followed by a single round of reduction to
remove the newly introduced terminal double bond. Alternatively,
an isomerase might act on 1-heptadecene to produce these 2- and
3-heptadecenes. Fatty acid analysis of M. producens 3L did not
reveal any unsaturated fatty acids with double bonds at these v-2
or v-3 positions, but instead, hexadecanoic acid, 11-hexadecenoic
acid, 9-hexadecenoic acid, and 11-octadecenoic were observed in
order of decreasing abundance (Figure 5). The lack of v-2 or v-3
unsaturated fatty acids in this cyanobacterium suggests that an
isomerase is likely involved in 2- and 3-heptadecene biosynthesis.
Consistent with this hypothesis, the required 16:0 fatty acid is the
most abundant fatty acid in M. producens 3L.
Phylogenetic DistributionThe phylogenetic distribution of the two known hydrocarbon
biosynthetic pathways among cyanobacteria is revealing of their
evolutionary history. The FAAR/ADO pathway is the most
widely distributed pathway taxonomically (122 of 139 strains) and
is therefore most likely to be the ancestral hydrocarbon pathway in
cyanobacteria (Figure 2). Alternatively the OLS pathway is only
found in a small number of cyanobacteria (17 of 139) and likely
evolved later than the FAAR/ADO pathway (Figure 2). A striking
finding revealed by this genomic investigation is the absence of the
alternative hydrocarbon pathway in a strain when either the OLS
or the FAAR/ADO pathway is present. More specifically, none of
the cyanobacteria with the OLS pathway possess the FAAR/ADO
pathway or any pseudogenes derived from it. In reciprocal fashion,
none of the organisms that possess the FAAR/ADO pathway also
possess the OLS pathway or any derived pseudogenes. Because the
OLS pathway contains many of the same enzymatic domains
found in PKS and fatty acid biosynthesis pathways (ACP, KS, AT,
KR, TE), homologs for each domain or even groupings of
domains can be found in all cyanobacteria. Thus, the distinctive
domains of this pathway are the fatty acyl ACP ligase (FAAL),
sulfotransferase (ST) and thioesterase (TE). These domains impart
the specificity of the pathway to first utilize fatty acid substrates,
activate the b-hydroxy fatty acid via sulfonation and then catalyze
hydrolysis and decarboxylation to produce terminal alkenes. While
homologous genes for each of these latter domains can be found
throughout cyanobacterial phylogeny, the unique domain archi-
tecture that makes up the OLS pathway is found only in OLS
containing cyanobacteria.
Hydrocarbon Pathway EvolutionMost of the strains that possess the OLS pathway are found
within a single clade on the 16S rRNA phylogentic tree (Clade A,
Figure 2). Cyanobacterial phylogenies using multi-locus sequence
analyses appear to exhibit the same placement of clade A within
the tree topology [48]. However, this clade also includes many
strains that possess the FAAR/ADO pathway, suggesting a unique
evolutionary history that may have involved multiple horizontal
gene transfer events. Thus, the evolutionary history of the OLS
pathway in clade A cannot be definitively attributed to horizontal
Figure 5. Fatty acid analysis of three cyanobacteria. Fatty acid analysis from single samples of Anabaena (Nostoc) sp. PCC 7120, M. producens3L, and Synechococcus sp. PCC 7002. All three strains exhibit similar proportions of hexadecanoic acid and 9-hexadecenoic acid; however, Anabaena(Nostoc) sp. PCC 7120 exhibits a higher proportion of 9-octadecenoic acid and 11-octadecenoic acid and tetradecanoic acid is absent. M. producens 3Lcontains 11-hexadecenoic acid and no octadecanoic acid. Synechococcus sp. PCC 7002 exhibited a similar composition to Anabaena (Nostoc) sp. PCC7120 yet has a higher proportion of hexadecanoic acid.doi:10.1371/journal.pone.0085140.g005
Cyanobacterial Hydrocarbons
PLOS ONE | www.plosone.org 9 January 2014 | Volume 9 | Issue 1 | e85140
gene transfer without additional information including further
genome sequences of OLS-containing cyanobacteria and phylo-
genomic comparisons.
The clade containing the Moorea strains (Clade B) as well as the
separate clade containing Leptolyngbya sp. PCC 6406 (Clade C) are
phylogenetically outside of this major clade of OLS containing
cyanobacteria. The Moorea strains, however, cannot be definitively
characterized as separate given the rather low bootstrap/posterior
probability support values. Nevertheless, the topological placement of
this group of filamentous tropical marine cyanobacteria in figure 2 is
consistent with previous phylogenetic analyses [24]. Additionally, the
OLS pathway found in the three genome sequenced strains of Moorea
and in Leptolyngbya sp. PCC 6406 is consistently separated into two
open reading frames with 9 to 17 bp intervening between the two
ORFs (Figure S4). This distinctive pathway architecture is not found
in any of the other 13 OLS-containing strains, and thus indicates a
potentially distinct evolutionary history for these two groups of OLS
pathways. Additionally, for all three Moorea strains and Leptolyngbya sp.
PCC 6406, the full AA sequences from the OLS pathway, as well as
the KS sequence considered separately, clade together and thus
suggest a common evolutionary history (Figure S5 and S6). Despite
the distinct pathway architectures and distinctive evolutionary
history, the resulting hydrocarbon products do not appear to be
fundamentally different as indicated by the production of 1-
heptadecene in Cyanothece sp. PCC 7822 and the three Moorea strains.
Leptolyngbya sp. PCC 6406 contains the OLS pathway, but
according to its 16S rRNA phylogeny, is very distantly related to
the other strains containing the OLS pathway (Figure 2). This
latter discrepancy between 16S rRNA and OLS gene tree
topologies suggests that the OLS pathway in Leptolyngbya sp. PCC
6406 may have been obtained by horizontal gene transfer.
Supporting this conclusion, the GC content of the Leptolyngbya sp.
PCC 6406 OLS pathway (64%) is higher than that of the entire
genome (55%); this contrasts with the OLS pathways in all other
cyanobacteria which exhibit similar GC contents to their
respective genomes (Table S3).
Thus, horizontal gene transfer appears to have played a role in
the evolutionary history of the OLS pathway in cyanobacteria;
however, the extent of this is unclear at this point. The competitive
exclusion of the OLS and the FAAR/ADO pathway is intriguing
and may suggest the presence of as yet unknown selective pressures
to maintain one or the other of these hydrocarbon biosynthetic
pathways, but not both. Further investigation of the physiological
and ecological role of cyanobacterial hydrocarbons as well as
further delineation of the phylogenetic distribution of these two
pathways may reveal insights as to the nature of this selection and
competitive exclusion pressure.
Supporting Information
Figure S1 Cyanobacterial 16S rRNA phylogeny andhydrocarbon pathway distribution for the compressedtree displayed in Figure 3. The 16S rRNA phylogeny is
displayed for the 20 cyanobacteria quantitatively characterized for
their hydrocarbon composition. Blue strain names indicate strains
possessing the FAAR/ADO pathway and red strain names
indicate those with the OLS pathway. Purple strain names
indicate a strain that does not have a genome sequence and
therefore the pathway is unknown. Cyanobacterial subdivisions
are labeled using colored branches following the key in the upper
left: Subdivision I. Uniceullular (Formerly Chroococcales), Subdi-
vision II. Baeocystous (Formerly Pleurocapsales), Subdivision III.
Filamentous (Formerly Oscillatoriales), Subdivision IV. Hetero-
cystous (Formerly Nostocales), Subdivision V. Ramified or True
Branching (Formerly Stigonematales).
(TIF)
Figure S2 Cyanobacterial 16S rRNA phylogeny andhydrocarbon pathway distribution for compressed treedisplayed in Figure 4. The 16S rRNA phylogeny is displayed
for the 32 cyanobacteria characterized for their hydrocarbon
composition as an overall percentage. Blue strain names indicate
strains possessing the FAAR/ADO pathway and red strain names
indicates those with the OLS pathway. Purple strain names
indicate a strain that does not have a genome sequence and
therefore the pathway is unknown. Cyanobacterial subdivisions
are labeled using colored branches following the key in the upper
left: Subdivision I. Uniceullular (Formerly Chroococcales), Subdi-
vision II. Baeocystous (Formerly Pleurocapsales), Subdivision III.
Filamentous (Formerly Oscillatoriales), Subdivision IV. Hetero-
cystous (Formerly Nostocales), Subdivision V. Ramified or True
Branching (Formerly Stigonematales).
(TIF)
Figure S3 Mass spectra of branched hydrocarbonsobserved in Fischerella sp. PCC 7414. The fragmentation
patterns were used to propose the locations of methyl group
substitutions on a heptadecane parent structure. In addition to the
fragment losses depicted, neutral ion losses were also present.
(TIF)
Figure S4 DNA alignment of all 17 known OLS pathwaysand CurM with annotated ORFs. Four of the strains have the
OLS pathway split into two ORFs. In these four strains the fatty
acyl ACP ligase is a separate ORF from the PKS portion of the
OLS pathway. Panel A shows all 17 aligned sequences with a red
square highlighting the break between ORFs for the four pathways
with two ORFs. Panel B shows the expanded red highlighted
region from panel A. CurM is a part of the curacin A biosynthetic
pathway which does not involve a FAAL [23].
(TIF)
Figure S5 Amino acid alignment and phylogenetic treeof all 17 OLS pathways and the CurM domain. All 17 of
the OLS pathways contain the same domain architecture. CurM
does not contain the FAAL and ACP1 domains. A maximum
likelihood tree is displayed on the left of the alignment to depict
the phylogenetic relationships between these pathways.
(TIF)
Figure S6 Phylogeny of the KS domain for 17 cyano-bacterial OLS pathways. The KS domains from Leptolyngbya
sp. PCC 6406 and the three Moorea strains clade together,
suggesting a common evolutionary history.
(TIFF)
Table S1 Summary of literature reports of cyanobac-terial hydrocarbon composition. Strains are organized by
subdivision on the left-hand side, and a ‘+’ symbol indicates that a
particular hydrocarbon was reported for this strain. The relevant
reference is listed in the right-hand column with full references
below. Findings that were validated using authentic standards or
established using derivitization techniques to verify hydrocarbon
structure are denoted with superscript 1.
(XLSX)
Table S2 Sequence and Pathway Information. Table of
genes used in this investigation (16S, OLS, FAAR, ADO) with
Genbank accession numbers and JGI IMG ID numbers.
NA = Not Applicable, UA = Unavailable, UNK = Unknown.
(XLSX)
Cyanobacterial Hydrocarbons
PLOS ONE | www.plosone.org 10 January 2014 | Volume 9 | Issue 1 | e85140
Table S3 Genome and OLS gene percent GC for all 17OLS pathway containing cyanobacteria. Genome status is